Techno-Thriller: Why Was Goldman Sachs So Worried About One Nerdy Sentence?

It sounds like the plot to a dozen movies: Picture a corporation so powerful that its tentacles circle the globe and reach into the highest corridors of power. Yet a single sentence on an ex-employee's obscure website forces it to move into action. That sentence is so important that it leaves the corporation with no choice but to make that employee ...

No, not disappear. They just made him delete it. (This is where the movie comparisons end.) But the question is, why? The sentence described the Goldman Sachs risk system, SecDB (which stands for securities database). It read: "Unbeknownst to most of the non-strategists, you could see basically every position and holding across the company, whether you were supposed to or not."

Without some digging we can't know that the sentence is true -- but why did it cause such a reaction? It was pretty well buried in a blog post by Antonio Garcia-Martinez, a former Goldman "quant" (financial analyst). The post is a long, kiss-and-tell piece about his reasons for joining Goldman, his experiences there, why he left, and why he's happier at his new start-up company. He says a number of unflattering things about Goldman Sachs in his post. So do a great many people, for that matter -- every day.

So why did Goldman Sachs bother making Garcia-Martinez delete this particular sentence, out of the reams of scandalous things said about them, and then contact Business Insider's Clusterstock blog (which had reprinted it) to deny that it was true? Because it could have a serious impact on Goldman's future.

First, consider the effect a revelation like this would have on Goldman's already-battered client relationships. The firm is struggling to overcome the now-public knowledge that it bet against some of those clients, causing them financial harm while claiming at the same time to "serve their needs." Personal relationships can influence a business deal even more than the corporation's reputation, which is probably one reason Goldman's still around. A corporate exec may continue to place his business with them even after he's heard the bad stuff, as long as he likes and trusts hiscontact there.

But what if our exec knew that his trusted "friend" at Goldman was aware of every position Goldman or its clients were taking against him (or had taken in the past), and that the guys betting against him knew instantly what he was doing? He might not want too much to do with his "friend" after that.

Then there's the question of market manipulation. Goldman has approximately 15% of the entire derivatives market. If its traders can immediately cross-check any deal they negotiate against what's happening across 15% of the market, in real-time, that could raise serious legal issues. And it could seriously undercut statements like these, in which Goldman's senior management tried to defend itself from accusations of fraud:

"We certainly did not know the future of the residential housing market in the first half of 2007 any more than we can predict the future of markets today. We also did not know whether the value of the instruments we sold would increase or decrease." (emphasis mine)

If they and their employees were tracking every deal in real-time they had a better picture of the those instruments' value than they let on. A database like the one Garcia-Martinez described, if it exists, would be an invaluable tool in getting a jump on the market -- or manipulating it.

But wait: it gets worse. David Viniar, Goldman's CFO, testified under oath to the Federal Crisis Inquiry Commission and claimed that Goldman didn't track its derivatives deals separately from its other transactions. FCIC panel chair Angelides pressed him: "Are you telling me you have no system at your company that tracks revenues or assets of contracts, and liabilities and payments under contracts? You have no management reports, no financial reports that track these contracts?"

"I've never seen one," Viniar answered. The Commissioners seemed to verge on accusing him of lying. "Nobody here really believes (that)," said one. Flash back to February of this year, when David Viniar said this in a presentation to investors: "Technology is fundamental to everything we do, from revenue-producing activities to enabling much of the control infrastructure of the firm." And here's another quote, from Goldman's Business Principles: "We take great pride in the professional quality of our work."

As the Wall Street Journal reported, Goldman has said that "credit trading desks ... are separated by industry group ... (and) traders are indifferent to whether they are selling clients a bond or a credit derivative." The Journal added: "The firm also said its technology systems firm-wide don't single out derivatives transactions."

Now comes the part of the movie where we place our sentence, that jigsaw puzzle piece, into context so that we get its full meaning:

" The Goldman Sachs risk system is called SecDB (securities database), and everything at Goldman that matters is run out of it. ... Database replication was near-instant, and pushing to production was two keystrokes. You pushed, and London and Tokyo saw the change as fast as your neighbor on the desk did (and yes, if you fucked things up, you got 4 AM phone calls from some British dude telling you to fix it). Regtests ran nightly, and no one could trade a model without thorough testing ... Unbeknownst to most of the non-strategists, you could see basically every position and holding across the company, whether you were supposed to or not. The whole thing was so good ..."

He's saying that Goldman Sachs has a first-rate, centralized data system that captures each deal in detail, and that everyone can see it as soon as it's posted. What's more, if I understand him correctly, he's saying that employees are required to run a detailed model of their deals before they can post them on the system. And all this information is stored on a database. How can a system can do all this and yet be unable to distinguish between a bond and a derivative?

Nevertheless, David Viniar testified under oath that Goldman's systems were so unsophisticated that he couldn't even tell the FCIC how much profit the company made from derivatives. Eventually, under continued pressure, Goldman provided the FCIC with an estimate which amounted to 25%-35% of its 2009 revenue. Yet Goldman is telling investors it won't lose any revenue as a result of the financial reform bill, and analysts believe them. "They've clearly seen the writing on the wall and are planning their moves ahead of time," said one.

That brings us to Goldman's plans to shut down its proprietary trading unit and spin it off into an independent hedge fund -- or move it into Goldman's asset management arm. Here's a question that probably hasn't been asked yet: Do they plan to use Goldman's SecDB, or any other Goldman systems, in that asset management firm? If this trading unit is moved into a hedge fund, will that fund 'rent' its computer systems from Goldman? Will all the traders and 'quants' at these various organizations be able to 'see' deals happening in real time? That could trigger calls for an SEC investigation or other actions to prevent Goldman from improperly using computer data. And it could raise questions about the other big banks' systems, too.

It's understandable why, given the implications of this nerdy sentence, Goldman would dispatch its publicists to tell Clusterstock it isn't true. As for Garcia-Martinez, he was asked why he deleted the sentence. "Prudence is the better part of valor," he answered -- followed, no doubt, by a click as a gloved hand placed the telephone back in its cradle.

Or course, this is no thriller. But the story raises serious questions, ones we should be asking anyway. If a bank is "too big to fail," its data is "big enough to misuse." It also shows why we need strong regulations behind the financial reform bill, to make sure that information doesn't become another "financial weapon of mass destruction." And it shows why we need to end "too big to fail."

The story also illustrates how much we don't know about what the big financial players are doing. It reminds us that we wont be able to protect the economy from future Wall Street crimes unless we keep investigating the ones that have already taken place.


Richard (RJ) Eskow, a consultant and writer (and former insurance/finance executive), is a Senior Fellow with the Campaign for America's Future. This post was produced as part of the Curbing Wall Street project. Richard also blogs at A Night Light.

He can be reached at ""